Atlas

Compare Models

Claude Fable 5.2 vs Grok 4.6

Claude Fable 5.2 leads by 10.2 capability points (≈95% interval +7.6 to +12.8): a clear lead over Grok 4.6 · P(higher) ≈ 100%

Wins 66 of 71 shared benchmarks (4 lost, 1 tied)

Capability vs. relative cost
Claude Fable 5.2Grok 4.6Closed weightsOpen weights

Each line runs through one model’s priced reasoning settings, cheapest to most expensive; grey points are other models’ priced settings in the same range. Results whose reasoning effort the source didn’t report are left out wherever a model has named effort levels. Capability and relative cost are the published estimates (relative cost is cost per task as a multiple of GPT-5 (high)); hover a setting, or use the arrow keys, to set it beside the other model’s same setting.

Claude Fable 5.2vsGrok 4.6capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins66 – 4 · 71 shared, 1 tied

Comparing matching setups only (3 benchmarks excluded) · suites counted once
← Claude Fable 5.2Grok 4.6 →

Claude Fable 5.2vsGrok 4.6wins by categoryData Nov 15, 2026Data: Artificial Analysis, Epoch AI, Arena and 19 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Grok 4.6 minus Claude Fable 5.2.

ARC-AGI-3 (Semi-Private)General QA81.22.1−79.1
FrontierMath Tier 4 (v2) - PrivateMath100.031.7−68.3
Humanity's Last Exam DiamondGeneral QA80.221.5−58.7
LMArena Agent Arena Net Improvement (Steering Burden)Agents30.4-0.2−57.7
Mystery Game PuzzlesGames90.034.0−56.0
ChartographyMultimodal73.420.5−52.9
AA-PersonalityBench Warmth indexChat & Writing82.931.5−51.9
Furniture Assembly BenchmarkMultimodal90.040.0−50.0
ProofBench v1.1Math100.051.0−49.0
AA-PersonalityBench Imagination indexChat & Writing67.620.0−48.1
Chess Puzzles (task version 2.0)Games79.033.0−46.0
SimpleQA Verified (Epoch anti-abstention)General QA94.049.3−44.7
FrontierSWE V2Code69.925.3−44.6
AA-PersonalityBench Organization indexChat & Writing54.314.3−40.4
HardGeoBench Avg DistanceMultimodal5521807−38.7
LMArena Agent Arena Praise vs ComplaintAgents73.4-2.5−38.0
Humanity's Last Exam (Text-Only)General QA74.342.9−31.4
GameBench 2Games99.168.7−30.4
FrontierMath Tiers 1-3 (v2) - PrivateMath95.166.0−29.1
AA-PersonalityBench Slop indexChat & Writing6.735.3+28.8
HardGeoBench <=10miMultimodal48.820.0−28.7
BioSecBench-Structure V1 (58 evaluations)Science43.714.9−28.7
LMArena Code Arena WebDevCode18361618−28.5
Bug Hunt BenchCode55.227.3−27.9
FrontierCode 2.0 (Artificial Analysis)Code47.619.8−27.8
HardGeoBench <=100miMultimodal60.032.5−27.5
ARC-AGI-2General QA93.867.1−26.7
HardGeoBench Country AccuracyMultimodal70.043.8−26.2
ChessBenchGames69.343.3−26.0
BuseyBench SVGMultimodal7.24.6−25.6
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine43.918.3−25.6
Conceptual Reasoning IndexIndexes79.962.5−23.9
HieroglyphBenchMultimodal43.620.3−23.3
HealthBench ProfessionalHealth & Medicine71.448.5−22.9
CritPt v2 (agentic, avg@5)Science84.962.3−22.6
LMCAGeneral QA70.648.5−22.2
MysteryMechanismScience51.830.6−21.2
CrossSearch (pass^3)Search56.035.0−21.0
ACCoRDGeneral QA89.471.6−17.8
Vibe Code Bench v1.1Code93.876.2−17.5
LMArena Agent Arena Steering BurdenAgents33.0-1.8−17.4
AA-PersonalityBench Conviction indexChat & Writing75.392.4+17.3
AA-PersonalityBench Expressiveness indexChat & Writing81.063.8−17.3
CritPt (avg@5)Science35.719.7−16.0
Epoch Capabilities Index (ECI)Indexes174158−13.2
Vals RSI Index v1.1 (four tasks)Agents37.925.5−12.4
LMArena Agent Arena Confirmed SuccessAgents22.0-2.6−12.3
BioMysteryBench (Vals Terminus 2)Science83.772.2−11.5
SciCode 1.0.1 (Artificial Analysis)Code67.756.5−11.2
ARC-AGI-1 Semi-PrivateGeneral QA98.587.5−11.0
AA-Briefcase v1.1 EloAgents17371540−10.9
APEX-Agents 1.1Professional Work87.677.3−10.3
APEX-Accounting 1.1Finance & Accounting70.460.3−10.1
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.745.7−10.0
AA-PersonalityBench Transparency indexChat & Writing81.071.4−9.6
Immigration BenchLegal74.066.0−8.0
DeepSWE v1.1Code74.366.7−7.6
MineBench — Bradley-Terry RatingGames19321811−7.3
AutomationBench-AA 1.0.6Agents74.267.0−7.2
Tax Agent BenchFinance & Accounting67.071.4+4.4
GDPval-AA v2.1Professional Work17511658−4.1
SWE-bench MultilingualCode96.692.7−3.9
MacroscopeBenchCode78.575.9−2.6
AA-LCR v1.1Search83.381.0−2.3
LMArena Agent Arena Bash ReliabilityAgents14.410.0−2.2
DAXBench v2Code88.386.2−2.1
AA-PersonalityBench Informativeness indexChat & Writing46.744.8−1.9
DTBenchGeneral QA99.297.3−1.9
OTIS Mock AIME 2024-2025Math99.799.2−0.6
GPQA DiamondGeneral QA93.994.0+0.1
LMArena Agent Arena Tool HallucinationAgents1.01.00.0

Claude Fable 5.2vsGrok 4.6normalised score marginsData Nov 15, 2026Data: Artificial Analysis, Epoch AI, Arena and 19 othersVillage LLM Atlasatlas.village.dev

Win counts use each benchmark’s declared score direction. Represented suite slices and the derived Capability score are excluded from headline totals. Missing or incompatible setups are excluded, never counted as losses.
Loading Atlas data…